Role of Diuretics in Cardiovascular Events and Mortality in Systolic Blood Pressure Intervention Trial A Post Hoc Analysis

Background In a post hoc analysis, we examined whether postrandomization diuretics use can explain and/or mediate the beneficial effects of intensive systolic BP lowering on cardiovascular disease and all-cause mortality in the Systolic Blood Pressure Intervention Trial (SPRINT). Methods SPRINT was a randomized, controlled trial of 9361 participants comparing the effects of intensive (systolic BP target <120 mm Hg) versus standard (systolic BP target <140 mm Hg) BP control on a primary composite cardiovascular end point in participants aged 50 years or older with systolic BP of 130-180 mm Hg. In time-varying multivariable Cox analyses, we assessed hazard ratios (HRs) of cardiovascular end points and all-cause mortality in participants on thiazide type, loop and/or potassium (K) sparing, or no diuretics. We also conducted mediation analysis to formally assess the role of diuretics in the effects of intensive systolic BP lowering. Results At baseline, diuretics were prescribed in 46% and 48% of participants in standard and intensive systolic BP-lowering groups, respectively, and in 46% and 74% in the corresponding groups during the trial. The lower risk of cardiovascular end points in the intensive group (HR, 0.75; 95% confidence interval [CI], 0.64 to 0.89) persisted after adjustment for postrandomization time-varying diuretics use (HR, 0.74; 95% CI, 0.62 to 0.89). Across the entire study population, time-varying diuretics use was not associated with cardiovascular end points (compared with no diuretics, HR for thiazide type, 0.89; 95% CI, 0.73 to 1.10, and loop/K sparing, 1.29; 95% CI, 0.97 to 1.73). However, thiazide-type diuretics were associated with lower risk of cardiovascular end points in the intensive (HR, 0.62; 95% CI, 0.46 to 0.85) but not in the standard (HR, 1.07; 95% CI, 0.82 to 1.39) group. In mediation analysis, HRs for total effect, direct effect (not mediated through diuretics use), and indirect effect (mediated through diuretics) of the intervention on cardiovascular end points were 0.66 (95% CI, 0.54 to 0.79), 0.67 (95% CI, 0.54 to 0.81), and 0.98 (95% CI, 0.88 to 1.10), respectively. The results were largely similar for all-cause mortality. Conclusions The favorable effects of intensive systolic BP lowering on cardiovascular end points and all-cause mortality in SPRINT were independent of and not mediated by time-varying diuretics use. However, thiazide-type diuretics use associated with benefit if intensive systolic BP lowering was targeted.